Designing High Performance Trays -The Art Of-

by

Ravi Marimootoo, Dr. Stefan Sander & Daniel R. Summers

Sulzer Chemtech Winterthur, Switzerland

Hans Nooijen & Peter Wilkinson

Shell Global Solutions, Amsterdam, Netherlands

Presented at

ERTC Petrochemical Conference Cannes, France, 29th September – 1st October 2008

Abstract

What is a tray? Some would only say a panel with perforations that provides a platform for

mass transfer; others would add, it is a piece of scientific art. It takes decades of experience to

understand the optimum working features of a tray through an understanding of the

application/service it is meant to work in, the efficiencies required to achieve the desired

output, and the craftsmanship needed to simply produce it. Through a look at the evolution of

trays from their simplest designs to their high performance counterparts, 2 case studies will be

presented to show the positive impact of high performance trays on process efficiencies and

on achieving mega throughputs under more stringent regulations and restrictions.

2

1. Introduction

When the engineer is faced with a seemingly simple question to describe a tray, very often,

the buzzing mind of the same engineer adventures into a paragraph of technical words that

would puzzle even the greatest experts about the topic. Trays come in all shapes and sizes

with various names such as, sieve, bubble cap, truncated, step-arc, multi-downcomer, mod-

arc, reverse flow, disc & donut, shed deck, ripple, etc,. These trays were all invented for

specific processes or services. For example we find use for the bubble cap and reverse flow

trays in low liquid load services while the disc & donut tray is well suited for highly fouling

services where efficiency is expected to be low. Understanding the benefits and working

limitations of a tray are key to identify the right internal for the right service, even more so,

when constraints such as plot space, lower impurity concentrations, and competition, push the

engineer to design better, more efficient, and cost-effective trays.

2. What is a Tray?

A tray is a circular disc, or part of, fixed in a distillation column of a chemical plant and

which provides a platform for an ascending vapour stream and a descending liquid stream to

interact. Theoretical plates have been established as a measure of separation efficiency in

distillation. Via experience, smart modeling, and rigorous calculations, theoretical plates are

then physically represented by trays. Trays are then designed and manufactured to bring to the

process the desired efficiency and capacity.

3. Tray Types

3.1. Conventional Trays

Among the conventional trays, a baffle tray does well to promote vapour and liquid

interaction, but due to its simplistic configuration Figure 1, separation efficiencies are at the

lowest. These trays are thus preferred in heat exchange applications and in fouling services

due to their open structure.

3

Figure 1 - Baffle Tray

A sieve tray, Figure 2 which consists of a perforated (holes) plate with compartment(s) known as

downcomers to accumulate liquid, shows good efficiency.

Figure 2 – Sieve Tray

With the increase in demand of chemical products to the market, plant owners have to

continuously adapt their existing designs to suit the market's growing needs. In essence, it

implies modifying existing equipment to handle higher throughputs while minimizing the

requirement for additional space or bigger equipment on the facilities. Licensors are often

under pressure to limit investment costs on new complexes or, in revamp cases, use existing

distillation column shells to fit higher performance internals instead of replacing the old with

new and bigger columns to handle the demand for extra capacity.

The invention of the V-Grid tray with its tapered, trapezoidal valves extruded from the tray

deck and oriented parallel to the liquid flow, allows for 5-10% extra capacity compared to the

sieve tray. The lateral vapour release Figure 3 of the V-Grid valves results in a lower froth height

4

than what can be typically observed with sieve holes. This, in turn, allows for more froth,

hence higher vapour and liquid loads, to fill up the space between trays before reaching the

flooding point.

Figure 3 – Lateral Vapour Release

The table below shows the performance of conventional trays for different fields of

applications in reference to a sieve tray.

Sieve Fixed Valve

Movable Valve

Bubble Cap Dual-Flow Baffle, Disc & Donut

Capacity Base Base Base Lower Much Higher Much Higher Efficiency Base Base Base Base Much Lower Much Lower Turndown 2:1 3:1 4:1 8:1 2:1 Not applicable Entrainment Base Lower Lower Base Base Base Pressure Drop

Base Base Base Higher Lower Much Lower

Maintenance Base Base Higher Much Higher Lower Lower Fouling Resistance

Base Higher Lower Much Lower Much Higher Much Higher

Corrosion Resistance

Base Base Lower Lower Higher (Scaling)

Higher

Application Low-High Flow Parameter

Low-High Flow Parameter & Fouling Services

Low-High Flow Parameter with high turndown requirements

Low flow parameter

Fouling Services

Fouling Services

Table 1 Conventional Tray Comparison

3.2. High Performance Trays

A high performance tray displays enhancements on tray decks and/or downcomers. The rule

of thumb is higher gas load implies enhanced active area and higher liquid load implies

enhanced downcomer area. The fine tuning between active areas and downcomer areas of a

5

high performance tray may at times require ingenuity and creative engineering founded on

strong experience when the software simulating the performance of a tray is bound by safe

correlations. Figures 4 and 5 below are illustrations of two types of high performance trays,

the Shell HiFiPlus tray and the VGPlus tray respectively.

Figure 4 – Shell HiFiPlus Tray

Figure 5 – VGPlus Tray

6

1

44

187

Feed

Vent

C3H8

1 ID=5.2m

ID=6.9m

54

Feed

55

178

177

51

4. China - C3 Splitter with Shell HiFi Trays

4.1 Background

In 2004, Sulzer delivered to a new petrochemical complex in China a set of Shell HiFiPlus

trays to be fitted in a C3 splitter Figure 6. The C3 splitter consists of 2 columns operating in

series with a side reboiler. The columns operate at high pressure > 20 bar(a) and utilize waste

heat (condensate) from the ethylene plant.

Figure 6 – PFD of C3 Splitter

4.2 Column Configuration

Tower A with an inside diameter of 6,900 mm has 187 Shell HiFiPlus trays and tower B, with

an inside diameter of 5,200 mm, has 51 Shell HiFiPlus trays. Due to the high number of

theoretical stages (150 – 200 in mega-throughput plants) required to achieve the desired

separation, C3 splitters tend to be tall. One goal is thus to minimize column height. This is

achieved by keeping tray spacing low although at the expense of comfort for the maintenance

crews. A low tray spacing, in turn, implies a reduced hydraulic capacity between two trays

since the froth created by the interaction of vapour and liquid has less space to settle in.

Tower B has a tray spacing of 400 mm. Tower A has tray spacings of 350 mm and 400 mm at

different sections. The feed to tower A is in the liquid phase and comes from two sources; the

bottoms of the deethanizer column and the overhead liquid from the Low Pressure

Depropanizer column. These two streams are combined before entering the tower.

Tower A

Tower B

7

0

10

20

30

40

50

60

70

80

0 50 100 150 200 250

Time (h)

Feed

rate

(t/h

)

0

100

200

300

400

500

600

700

800

900

1000

Ref

lux

/ Int

erna

l rat

es (t

/h)

In August 2006, a test run was carried out on the C3 splitter which operated with 98% feed

rate but higher reflux ratio. This brought the hydraulic operation of the column close to

design. The test run consisted in monitoring the column under steady state Figure 7 for 48 hours.

Figure 7 – Steady State Run August 2006

With a feed rate of 60,000 kg/hr for 550 kg/hr as vent and 55,810 kg/hr as product, the

material balance seems too perfect. However, this is brought to acceptable levels when

considering the standard deviation of 1,840 kg/hr (3%) observed in the product. With the data

collected, Sulzer was able to run its own simulations with its proprietary ProII thermodynamic

model to determine internal loads and tray efficiency of the tower.

Table 2 below compares the design basis with the results obtained from the test run data

collected in August 2006.

Design Basis Test Run Results Propylene Top Purity > 99.5% 99.83% Propylene in Bottom < 3.5% 2.73% Pressure Drop per tray 5.5 mbar 5.2 mbar Maximum Useful Capacity 71% 72% Tray Efficiency 85% 90 – 95% Table 2 Design basis versus test run results

8

Key aspects that were considered in the design of the Shell HiFiPlus trays are summarized

below:

� Track record of existing C3 splitters equipped with Shell HiFiPlus trays was necessary

for consistency and design optimization.

� Minimum tray spacing without sacrificing on useful capacity.

� Setting the optimum net free area based on statistical references of existing C3

splitters with similar operating conditions. Net free area has shown to have an impact

on efficiency.

� Flowpath length: With a multi-downcomer tray, flowpath length becomes an

important parameter as this affects the efficiency of the tray. Too short a flowpath

length can result in loss of efficiency due to too little contact time between vapour and

liquid on the active area.

4.3 Conclusion

Although the set of Shell HiFiPlus trays were delivered for a brand new Chinese

petrochemical complex, the C3 splitter, which requires a large number of theoretical stages,

needed to be equipped with high performance trays to minimize column height and inside

diameter while still ensuring useful capacity and process efficiencies were met according to

design.

5. Russia - C3 Splitter with VGPlus Trays

5.1 Background

On the 12th December 2006, Sulzer entered into contract with a Russian Olefins producer to

revamp their C3 splitter with high performance trays. The objective of the client is to go into a

2-phase upgrade from a 300 KTA plant to step-wise new capacities of 360 KTA and 430

KTA respectively. Accordingly, the C3 splitter would see an increase in feedstock capacity

from 20 tons/hour to 30 tons/hour. Furthermore the revamp relating to engineering, delivery

and installation was to be carried out over a span of 6 months.

9

5.2 Column Configuration and description

The C3 splitter Figure 8 consists of two column shells in series (Tower 1 and Tower 2). The two

column shells are effectively operating as a single column: overhead vapor from the first

column is sent to the bottom of the second column and liquid from the sump of the second

column is pumped back to the top of the first column. The overhead pressure of the second

column is approximately 9.81 barg. Polymer grade propylene is obtained as overhead vapour

from the second column. (Recycle) propane is obtained as bottom liquid from the first

column. The reboiler and condenser of the C3 splitter are integrated by means of a heat pump.

Overhead propylene vapour is compressed to 15.30 barg and condensed in the internal

reboiler/condenser. An auxiliary reboiler and auxiliary condenser are present to account for

the net remaining duty. The reflux is sub-cooled by vaporizing the feed to the column. Each

column has an inside diameter of 3,400 mm. Tower 1 was equipped with 96 sieve trays and

tower 2 with 106 sieve trays and both columns had a tray spacing of 550 mm.

Figure 8 – PFD of existing C3 splitter configuration

10

Feed rates and design basis were submitted as follows.

Table 3 Technical Specification

Based on the information provided, Sulzer carried out simulations using ProII and its

proprietary thermodynamic model. Guarantees were required on top propylene purity,

propylene quantity in the bottoms, capacity, pressure drop and mechanical integrity. A

sensitivity analysis Figure 9 was essential to decide if more trays should be fitted to increase the

number of theoretical stages or whether a 1-for-1 revamp would be sufficient to meet the

process and hydraulic requirements. A 1-for-1 revamp also meant re-using existing

attachments thereby minimizing installation time. The sensitivity analysis revealed that with

the existing condenser, the amount of reflux that could be sent to the column would be

sufficient to ensure that a 1-for-1 revamp to high performance trays would achieve the targets.

11

Reflux vs NTS

250

260

270

280

290

300

310

320

330

340

350

360

140 150 160 170 180 190 200 210 220

NTS (-) incl Reboiler and Condenser

Ref

lux

Rat

e (to

/hr)

Figure 9 – Sensitivity Analysis

Since there were no requirements to fit additional trays, which implies decreasing tray

spacing, the choice of VGPlus trays was deemed suitable for this revamp. It is important to

study carefully the requirements of a revamp. Chordal downcomer trays are accepted by the

industry to show higher efficiencies than multi downcomer trays due to the longer flowpath

lengths available. However, if more trays are required in a column to increase theoretical

stages probably due to limitations with the existing reflux condenser, multi-downcomer trays

such as the Shell HiFi trays are suitable as they can be fitted for tray spacing as low as 300

mm. With a tray spacing of 550 mm available and a 1-for-1 revamp, a set of VGPlus trays

with mod-arc downcomers were adapted to the existing attachments. It should be pointed out

that the client did not allow welding to the column wall. As such, centre and off-centre

downcomers were enveloped and adapted to the support rings without any welding required

to the column wall. The VGPlus trays Figure 10, after engineering, were presented as "FINAL"

to the client as shown below.

12

Figure 10 – VGPlus Tray Detailed Drawing The high performance trays were successfully installed within the shut-down window in June

2007 and in August 2007, the first set of operation data were made available to Sulzer for

analysis. Due to high fluctuations in feed quality varying between 60% to 80% propylene and

the planned revamp of remaining columns in this plant, the operators are running the C3

splitter at 20 tons per hour of feed. From the operation data submitted after the revamp of this

C3 splitter, the following table gives an insight into the results observed with the VGPlus

trays.

Design Basis (30 t/hr feed) Test Run Results (20 t/hr feed Propylene Top Purity > 99.8% 99.92% Propylene in Bottom < 3.0% 0.86% Pressure Drop per tray 6.4 mbar 5.5 mbar Maximum Useful Capacity 88% 70% Tray Efficiency 93% ~ 95% Table 4 Design basis versus test run results

5.3 Conclusion Although the Russian C3 Splitter is not currently running at the capacity for which the trays

were designed for, the results obtained from the test run data gives confidence that with the

reserves accounted for in the simulation and hydraulic design, the column will perform

according to design when switched to maximum capacity. The message to be taken out of this

project is that the choice of a higher performance tray is not solely dependent on its ability to

meet process requirements but also on indirect factors such as welding constraints on the

column shell, shut-down time, auxiliary capacities, and investment levels.

13

References

1. Loris Tonon, Pietro Bomio, Wouter Ankersmit, and Jose Bravo, "Debottlenecking &

Energy Savings in Ethylene Plants" ARTC Petrochemical, Kuala Lumpur Malaysia 12-14th

February 2001.

2. Markus Fischer Sulzer T-HbT T "Technical Handbook Tray Columns Revision 1.2" 28th

February 2007.

3. "Shell Design And Engineering Practice DEP 31.20.20.31-Gen" September 2002.

4. G.X. Chen "Sieve Tray Efficiencies" Fractionation Research Inc. FRI Topical Report No.

126, 1st May 1997

5. Tony J. Cai "Valve Tray Hydraulic Study" Fractionation Research Inc. FRI Topical Report

No. 137S, 31st May 2002

6. Daniel R. Summers, Darcy Ehmann, "Enhanced V-Grid Trays Increase Column

Performance" AICHE Annual Meeting, Indianapolis, Indiana November 2002.